How to remove the Sky from the Point Cloud using the Annotation Tool

In aerial oblique or terrestrial projects, part of the images is covered by sky. The sky does not consist of meaningful keypoints; however, it can happen that some of the sky keypoints extracted are matched in the images and reconstructed in the 3D model introducing noise to the model. It is possible to annotate the sky not to take it into account for processing.

To remove the sky from the Point Cloud:

1. Click View > rayCloud.
2. In the section Layers of the left sidebar, select one image that is covered by sky.
3. In the right sidebar, click the Annotation tool button pencil.png.
4. In Annotate, select Carve.
5. Annotate the sky in the image, following 202560549 (steps 7-13).

 
Important:
  • If step 2. Point Cloud and Mesh is already processed, the effect of the annotation is immediately visible in the 3D View of the rayCloud, but the results saved in the results folder are not updated.
  • If step 2. Point Cloud and Mesh is not processed, the annotation will affect the point cloud displayed in the rayCloud after processing and the results saved in the results folder. 
 
Important: The Carve annotation is not "forgiving"; i.e. all the 3D points that are located on the rays connecting the camera center and the annotated pixels are deleted. For more information: 212262883

6. Repeat steps 2-5:

    • If step 2 is already processed, till the sky points are removed from the 3D View of the rayCloud.
    • If step 2 is not processed yet, till the sky is annotated in all the images that it appears. 

7. Make sure that the Use Annotation box is selected in the Processing Options202560549 (step 14).
8. Process step 2. Point Cloud and Mesh to remove the annotated pixels from all the results.

 
Tip: Since the Carve annotation is not "forgiving", if some pixels are wrongly annotated, the Point Cloud will be affected.
It is recommended to annotate the images once step 2. Point Cloud and Mesh is processed. The effect of the annotation is immediately visible in the 3D View of the rayCloud and it is easy to notice if some pixels were erroneously annotated so as to be corrected immediately. Once the annotation is completed, step 2 needs to be processed again, to update all the results saved in the results folder. 
 
Was this article helpful?
0 out of 0 found this helpful